My Religion is "Something Else"

Measuring religion is hard, Chapter #20951

One of the organizing principles of my public-facing work is a simple one: make difficult things look easy. The vast majority of my audience doesn’t want to read an essay on whether probability-based sampling is worth the additional cost compared to opt-in sampling. Most of you aren’t interested in 2,000 words on question-ordering effects or how tweaking response options can lead to vastly different results. In other words, showing how the sausage is made isn't exactly fun.

But, indulge me just this one time because I think that a lot of you will find this interesting and informative. A growing number of Americans just don’t know what basic religious words mean. A couple of years ago, I published an article in the Journal for the Scientific Study of Religion called “The consequences of response options: Including both ‘Protestant’ and ‘Christian’ on surveys”, which I turned into a piece for Christianity Today.

Here’s the central point of that article: young people don’t know what the word Protestant means anymore, and that’s going to cause major problems in measuring religion going forward. The Nationscape survey (with a sample size of 477,000 respondents) asked about religious affiliation. The first five response options were:

• Protestant

• Catholic

• Latter-day Saints

• Orthodox (Russian, Greek, Eastern)

• Christian, other than above

I calculated the share who chose the ‘Protestant’ option and the ‘Christian’ option by age. Here’s what that looks like.

As you can clearly see, we have a major problem. Among the youngest adults in the sample, almost none of them selected Protestant. In fact, fewer than 10% did so until you get into respondents in their late 30s. In contrast, large numbers simply said they were Christian—at least 20% of those in their late teens and early 20s.

Once you look at those age 40 and older, the Protestant responses rise rapidly. It’s 20% among people in their mid-50s, climbs to about 30% in the early 60s, and reaches over 40% among the oldest adults—while only about 5% of them select just Christian.

The upshot: unless we start including the Just Christian option on surveys, we are likely undercounting Protestants in a significant way.

But there’s another dimension to this measurement debate I want to dig into today: the dreaded “Something else” option.

Here’s how it works. Everyone is asked: “What is your present religion, if any?” Then they are given these response options:

1. Protestant

2. Roman Catholic

3. Mormon

4. Eastern or Greek Orthodox

5. Jewish

6. Muslim

7. Buddhist

8. Hindu

9. Atheist

10. Agnostic

11. Nothing in particular

12. Something else

If someone checks “Something else,” they are directed to a text box where they can type in a description of their religion. Let me tell you: that’s a black hole of nonsense that makes me want to microwave my hard drives and live in the woods. But now I finally have a reason to poke around in that part of the data. Caroline Soler published a file that helps parse through those free response options and tries to classify them into their appropriate categories.

This is just the share of the sample that chose the “Something Else” option in the Cooperative Election Study in each year since 2010.

For a while, it hovered around 5% of the sample. It bounced up and down but stayed near that level. Then in 2020, it began to rise. Between 2020 and 2023, it ranged from 7% to 7.5%. In the most recent survey (late 2024), 8.4% of respondents said their religion was “something else.”

That’s a huge number when put in context. Add together the shares of LDS, Orthodox Christians, Jews, Muslims, Buddhists, and Hindus—that’s only about 6% of America. The best estimate for atheists is 7%. Mainline Protestants are now around 9.5%. So, 8% of U.S. adults saying something else is a big deal.

What makes someone more likely to pick “something else”? I looked at four possible factors using the most recent data: education, race, age, and political partisanship.

It’s probably no surprise that those with the lowest levels of education are the most likely to pick something else—14% of people without a high school diploma did. But as education increases, so does the likelihood of choosing a traditional religious label. Among college graduates, just 5% selected something else.

In terms of race, white and Asian respondents were most likely to identify with a traditional category, while 12% of Black respondents and 10% of Hispanic respondents picked something else.

I’ll admit: the age results surprised me. The group most likely to select something else were those between ages 55 and 64. Young adults were more likely to choose a conventional label—but the gap is small, just two percentage points.

Finally, let’s talk about partisanship. One thing jumps out: Independents. They are far more likely than Democrats or Republicans to pick something else. I’ve written before about people who are “dropping out of everything”—religion, party identification, ideology. This looks like another sign of that trend.

Okay, so if we actually sift through all the nonsense of the free responses and try to sort them back into the traditional categories, what do we end up with? Well, as previously mentioned, Caroline Soler tried to make that happen when she published a dataset recently.

She basically just combed through all those text answers and tried to figure out if they just didn’t fully understand their tradition. For instance, if someone wrote “Baptist” in the box, that would be reclassified as Protestant. She did this for every dataset from 2016 through 2024. A total of more than 17,000 responses.

When all those answers are parsed and recoded, what you get is a bit disappointing - about two-thirds of them are still in the “something else” bucket. In other words, it’s just not possible to easily sort them into our standard religious classification scheme. There are a bunch of reasons for this - researchers want to be extra cautious in putting someone in a category unless they are completely sure that it’s the correct one. The “Baptist” respondent is an example of a clear case. Most aren’t nearly so black and white.

Take this one for instance, “Christian (I don't like the term Protestant).” Do we actually call that person a Protestant when they expressly reject that label while also attending a Protestant Church? As you can see, almost all the recodes did end up sorting people into the Protestant category, though. Which speaks to the problem I mentioned earlier - people just don’t know what the word Protestant means.

So, among those who eventually got reincorporated into the Protestant classification, what actual words did they type into that box? I tried to parse that as best as I could. I removed all the capital letters so it would be easier to match cases. But I didn’t fix the spelling at all. Which, as you can tell, is a real problem when trying to use computer code to put people in the same category.

Far and away the most popular response among the “something else” Protestants was that they were a Christian. This was the case for over 3200 of the 5500 who eventually got reclassified. The next most popular option was “Baptist” followed by “non denominational.” Then there were a bunch of traditional Protestant families like Methodists, Lutherans, and Pentecostals.

Just for fun, I pulled out a handful of my most favorite free responses in the survey.

1. relationship with Jesus Christ. NOT a religion

2. Personal relationship with Jesus Christ

3. Just an Ordinary Christain

4. JESUS HOLY SPIRIT filled GOD loving CHRISTIAN

5. I grew up going to the catholic and the episcopal church because my moms side was catholic and my dads side was episcopal.

6. Christian, Non affiliated with a church but I go by what IS WRITTEN

7. Christian that believes you need to be born again but I'm not no evangelical.

8. A little atheist and Christian still trying to figure it out

9. I gave up on Christianity and more or less became an atheist, but instead of putting false hope into an all knowing divine being, who’s tenants are convoluted, I now simply believe that “something” is watching us beyond the stars. They slowly turn the wheel of fate and are neither good nor evil.

10. You should know never ask questions about religion or politics

If you would like to check out the full list, it’s right here. Feel free to drop some great ones in the comments, too.

I know that this post is written in a pretty light hearted way, but behind that is a very real and growing problem - we are trying to put people into categories that they don’t fully comprehend. With the rise of anti-institutional Christianity (with non-denoms) and all this remixed religion (yoga, ayahuasca), it’s not going to get easier.

But changing up question wording and response options is not something to be done lightly. We still need comparability. If we make big changes, we’ve essentially lost the ability to make simple statements about the size and trajectory of American Christianity.

Welp. There is it. A peek behind the curtain. But you can count on me to continue trying to make it easier to understand American religion.

Code for this post can be found here.

---

Ryan P. Burge is a professor of practice at the Danforth Center on Religion and Politics at Washington University.